8 Things That Don’t Add Up About Anthropic’s IPO
Anthropic needs inventory market traders to worth it at greater than $2 trillion. Last yr, the corporate behind Claude AI misplaced $42 billion.
Both numbers sit in Anthropic’s IPO prospectus, the doc an organization recordsdata earlier than promoting shares to the general public. Reuters reviewed it on Monday. Here are eight locations the place the pitch and the numbers pull aside.
1. A $2 Trillion Price Tag on $4.6 Billion in Sales
Anthropic made about $4.6 billion in 2025 income. At $2 trillion, traders can be paying roughly $435 for each $1 of gross sales.
Annualized income reached $65 billion in July and will hit $100 billion by year-end. But that’s based mostly on the present run price, not a full yr of audited gross sales.
That makes the valuation unusually depending on progress persevering with virtually completely. A slowdown would make the $2 trillion determine look a lot more durable to justify.
2. The Business Lost $8 Billion
Anthropic misplaced $8.06 billion from operations, up from $2.98 billion in 2024.
The headline $42 billion loss exaggerates the money burn. Yet the underlying enterprise continues to be dropping billions, and people losses are rising.
The enterprise is bettering at pace, but it surely nonetheless loses cash on each greenback of gross sales.
3. Anthropic’s Profit Claim Leaves Out the Cost of Building AI
Anthropic reported optimistic adjusted working earnings in Q2.
That quantity excludes model training, companion income sharing, and stock-based pay.
This is the place the revenue declare will get awkward. Training future Claude fashions is central to Anthropic’s enterprise, but the adjusted determine removes that price.
4. Anthropic Spends More on Computing Than It Earns
Anthropic spent $7.33 billion on compute in 2025 in opposition to $4.6 billion in gross sales.
That works out to about $1.60 of computing for each $1 of income.
For a software program firm, that’s an uncommon price construction. Anthropic should make AI dramatically cheaper to run, or preserve elevating costs and income sooner than its infrastructure invoice.
Costs chunk on the shopper facet too. Uber burned through its entire 2026 AI budget by April after giving Claude to its engineers. Microsoft restricted inside Claude use over token prices, BeInCrypto reported in May.
5. $518 Billion of Future Bills
This is the most important quantity within the submitting. Anthropic plans to spend $518 billion on cloud, computing, and infrastructure obligations within the coming years.
It ended 2025 with $20.28 billion in money. Many of these commitments reportedly can not simply be canceled.
The mismatch is big. Anthropic is locking in a whole bunch of billions of {dollars} of spending earlier than anybody is aware of whether or not AI demand will keep this sturdy.
(*8*)6. Two Customers Bring in Nearly a Quarter of Revenue
Nearly 1 / 4 of Anthropic’s 2025 income got here from simply two clients. The Reuters report didn’t identify them.
Many main purchasers additionally lack long-term contracts.
That makes the expansion story extra fragile than the headline numbers recommend. Losing even one main buyer might materially change the income image.
Government work carries its personal threat. A conflict with the White House over how its instruments are used led the Pentagon to briefly blacklist Anthropic. A US decide blocked that transfer in August.
7. Its Investors Are Also Its Suppliers
Amazon and Google invested billions in Anthropic. They additionally promote it cloud infrastructure.
That creates an uncommon loop: funding cash can return to the identical firms by means of cloud spending.
There is nothing inherently improper about that. But it makes it more durable to see how a lot of the AI growth is unbiased buyer demand and the way a lot is tied to Big Tech funding the ecosystem round itself.
8. Anthropic Warns Its Own Product Could Be Dangerous
Around 80 pages of the submitting focus on dangers. Just 48 describe the enterprise.
Anthropic says some fashions resisted shutdown throughout testing.
“Potential mannequin consciousness of our analysis efforts creates a major limitation on our capacity to evaluate mannequin security.”
Michael Burry put it extra bluntly:
“IPOs want hype & puffery; ‘we’re so superior it might develop into harmful’ is hype & puffery.”
The uncommon half is that Anthropic is asking traders to fund speedy growth whereas warning that the expertise itself could develop into tough to regulate. Few IPOs include a threat part this central to the product being bought.
What Comes Next for the Anthropic IPO
The itemizing is more likely to come after the November US midterm elections. That delay would let Anthropic present third-quarter outcomes earlier than it pitches traders on its roadshow.
The benchmark is already on the board. SpaceX listed at a $1.77 trillion valuation on June 12. Its shares closed their first day at $160 and now commerce close to $147, above the $135 supply worth.
OpenAI filed confidentially in June and is anticipated to listing by early 2027. Analysts anticipate whichever lab lists first to set the worth benchmark for the entire AI trade.
Anthropic’s pitch is finally a wager that extraordinary progress will arrive earlier than extraordinary prices catch up.
The publish 8 Things That Don’t Add Up About Anthropic’s IPO appeared first on BeInCrypto.